Never Mind The Hype Cycle
It explains nothing about AI or any other technology; it's just human behavior
Ah look…
There it is again: The chart that’s been rolled out more times than Nickelback at a halftime show.
It is of course the Gartner Technology Hype Cycle, enjoying its most talked-about showing since its wildly controversial “Crypto” world tour of 2022.
But while the graph stays the same, the narratives applied to it by writers and LinkedIn Hot Takers vary widely. To some, AI’s adherence to the Hype Cycle is proof that AI will definitely be transformational, we are just too early. To others, it’s just another get-out-of-jail card made by the tech bro establishment to excuse a pathological tendency to over sell and under-deliver.
So what should YOU make of it? Is AI just following the cycle of hype that Gartner presciently predicted all technology would do? Or does AI represent some exception to this norm?
In order to see the hype and hysteria for what they are, it is helpful to see the hype cycle for what it really is.
It’s much simpler than they want you to think.
It’s Not A Technology Hype Cycle, It’s A Human One
Amid the hype and hysteria, some quality, balanced opinions have been published, including several on Substack.
recently agreed that AI had passed peak-hype, but cautioned readers against writing it off. acknowledged the change in sentiment toward AI, but highlighted that we often rush to use technology in the wrong ways, creating discontent. shared a simpler model from Tyler Cowen that describes just 3 steps of Hype, Calm and Revolution.
These writers succeed by not delegating their brain to Gartner, and rather than asking where we are in Gartner’s hype cycle, they reflect on what the real driving forces are and how to interpret them.
These are great, but to me these all seem to circle but not quite a simpler truth that much more effectively makes sense of what is happening.
The Hype Cycle is not actually a technology phenomenon at all,
It’s a human one.
That’s how Gartner gets away with every technology fitting neatly to their model. It’s because EVERYTHING novel and exciting fits into this model!
Here, I’ll prove it to you…
The Human Hype Cycle
Suppose you first saw the hype cycle it was titled not as a technology phenomenon, but as one of these:
Or this
(You can fill in the blanks of events depending on your team sport of choice, but this certainly works for Arsenal fans)
Or how about this:
Or how about - for those of us that write long form articles like this one:
This trend line would also describe your experience of your dream holiday, that fancy car you think you want, or even that new internal corporate strategy that every employee is totally going to get behind.
In human terms, the stages look like:
1: New Possibilities
You encounter something very exciting, that represents a sharp change from before.
He/she is nothing like your previous partners - OMG they could be THE ONE?!?
That new player is going to score 100 goals this season and win you the league.
Your new job is finally the opportunity you’ve deserved, with none of the internal politics, (or that dick Kevin from accounts that had it in for you)
This. Changes. Everything.
2: The Climax
You get married. A Dream honeymoon, you can barely wipe the smile off your face.
You smash that first big presentation at work, the player is top scorer for the league… you get the idea.
Until…
3: Reality Bites
It turns out Mr/Mrs perfect is actually kind of annoying to live with. Their habits are starting to grate on you, they get angry when you go golfing/hang with friends/work late.
This isn’t what I expected.
Oh, and Mr Superstar Striker hasn’t scored for a month now. (How much did we pay for him? What a waste of money!)
And that guy Kyle in accounts - he’s an even bigger dick than Kevin ever was.
Ugh. This isn’t what I expected.
This. Changed. Nothing.
4: Reconcile and Respect
It can take us humans some time to realize the true value of an innovation, and it is often not valuable in the way we first thought.
(Just ask the Mesopotamians who allegedly used the wheel for pottery for 500 years before realizing it would be cool running on its side, under a cart)
Maybe your love didn’t turn out as you expected on your first date, but they have other qualities and virtues that you didn’t expect, but now respect and love them for.
And that superstar striker? Turns out he’s actually a solid holding-midfielder, and has a ton of assists. Sure, he hasn’t scored as many as you’d hoped, but he’s a great mentor and we are a better team when he’s on the pitch.
We changed. I grew.
In short, when humans encounter something exciting, we get excited. Sometimes over-excited.
(Especially when tech leaders over-inflate expectations to inflate their valuations, but there are enough posts on that already.)
So it really all just comes down to this:
Happiness = Reality - Expectations
As long as reality exceeds expectations, we are pleasantly surprised. We are Happy. Maybe even delighted.
But when expectations are higher than reality can deliver, we get disappointed and lash back.
(Or just create amusing videos about it)
Strip away the fancy labels of Gartner’s curve and it’s easy to see that’s all that’s really happening here: The gaps between what we expect, and the reality of AI’s capability.
This holds true for the previously mentioned human examples, as well as that of technology.
Your partner, the new player, or you in a new job all get a brief honeymoon period where initial excitement masks reality, but mostly it takes time really grow into a role and into expectations.
A simpler way to appraise AI maturity
By simplifying the Hype Cycle to simply expectations vs reality, 3 possible states emerge for applications to move through.
When deciding whether a product or use case is worthy of your investment - whether that’s as a customer, a founder or investor - try seeing if it fits one of these states:
1: Reality Meets Expectation
AI doesn’t need to be fully sentient to fulfill its promises of utility. It can still offer incredible value when AI's current capabilities align with or exceed user expectations.
Here are some examples of use cases where AI is already meeting or exceeding expectations:
Answer Engines like Perplexity that solve the hallucination/fact problem by providing inline reference for immediate fact checking
First Draft Assistants like ProcureSpark for RFPs that automate mundane and time consuming tasks to augment human labor, rather than attempt to replace it
Entertainment like using Luma Labs Dream Machine to animate old family photos for a fun and remarkable alternative to a birthday card, where absolute accuracy and believability is not required
2: Reality Demands Augmentation
If AI is almost ready to meet expectations for a given application, but not quite, is it possible to use humans to close the gap and meet expectations now?
For example, DXM focuses on the problem of brand consistency in AI generated marketing assets, which is paramount for marketers, but challenging for AI to deliver. This can be solved by a high level of setup and training, but is more than most marketers would undertake. DXM addresses this by employing human AI artists to do that for them.
3: Reality Isn’t Ready
If the gap between AI’s capability and customer expectations is too large to be augmented by humans, there are two choices
Proceed anyway and hope capability catches up or users lower expectations
Wait until it is the right time
Few of us would deliberately choose option 1, but I propose that it is at the root of most of the applications that have received unsympathetic scorn as they have failed.
Hear me speak more about these examples in this recent interview for AI Distilled
See Through The Hype And The Hatred
Public Enemy were right when they said “Don’t Believe The Hype” as were The Sex Pistols when they said “Nevermind The Bollocks”
The drama of both the speculation and the backlash create a smoke screen that the thinking person needs to see through to understand the true value of current and future applications. Having a frame of reference like Gartner’s Hype Cycle is helpful when it can normalize seemingly exceptional trends, but not when the this model is taken as proof in itself of a particular viewpoint.
Hype occurs not in technology, but in humanity.
It’s not that all technologies obediently follow a path that Gartner laid out for them. It’s that humans predictably get over excited about novel things, and pissed when reality fails to meet expectations. So the most helpful lens through which to view technical applications is in how it maps to human expectations.
I’ve found this framing of expectations vs reality to be helpful in sorting the valuable applications from the wasteful ones.
I hope it is valuable for you too.
PS. You can see my full interview with AI Distilled on YouTube where I share some thoughts on:
AI and The Hype Cycle
The Impact of AI Answer Engine and AI Agents on marketing
Some of the AI products represented within the
community that I founded and continue to run.
Wait, so you're saying the Gartner Hype Cycle itself is currently at the "Trough Of Disillusionment" stage?
Kidding, kidding.
Thanks for the shoutout!
As a former Gartner analyst, I support this assessment.